RE: [Algorithms] (Multi Layer Terrain) Triangle strips for "br oken" square patches
Brought to you by:
vexxed72
From: Cem C. <cce...@nv...> - 2004-02-17 20:51:27
|
About NvTriStrip: I had actually fixed these memory leaks some time ago, but the change never made it out to the website. Sorry about that. It's now been updated with the latest version: http://developer.nvidia.com/object/nvtristrip_library.html See the "NVTriStrip Library with Source" link. thanks, -Cem -- Cem (pronounced "Jem") Cebenoyan Software Engineer, Developer Relations NVIDIA Corporation > -----Original Message----- > From: Alen Ladavac [mailto:ale...@cr...] > Sent: Sunday, February 15, 2004 1:38 AM > To: gda...@li... > Subject: Re: [Algorithms] (Multi Layer Terrain) Triangle > strips for "broken" square patches > > > We did report both the leaks and the performance. They said > they'll fix it. But in the mean time, I thought it would be > useful to let people know it leaks memory, so they don't have > to find out the hard way (like we did :/). > > I don't think we'll go with fixing the source code for > nvtristrip ourselves (especially since it uses stl - sic! :) ). > > But, I've been thinking of this idea of using just > cache-coherency improvements, instead of full stripping. Now > that it was mentioned again, it got me thinking of it again. > I recently found a paper titled "Universal Rendering > Sequences for Transparent Vertex Caching of Progressive > Meshes" by Alexander Bogomjakov and Craig Gotsman. > (http://www.cs.technion.ac.il/~gotsman/caching/) It is very > similar to what Tom Forsyth suggested with space-filling > curves, just that this works for general meshes. Has anyone > tried something like that? > > Alen > > > > ----- Original Message ----- > From: "Gary McTaggart" <ga...@va...> > To: <gda...@li...> > Sent: Friday, February 13, 2004 19:05 > Subject: RE: [Algorithms] (Multi Layer Terrain) Triangle > strips for "broken" square patches > > > We use NVTristrip for preprocessing to generate cache > friendly trilists. I wouldn't spend much more effort on > tristripping than that if you can live with the time that it > takes to process your data. If there are leaks in their > code, I believe that you can download the source code for it, > or better yet, talk to Nvidia and get them to fix it. They > are pretty good about fixing their tools if they are broken. > Of course, if you are trying to strip at runtime, then you'll > probably want to find something else. > > Gary > > > -----Original Message----- > > From: gda...@li... > > [mailto:gda...@li...] On Behalf Of > > Ivan-Assen Ivanov > > Sent: Friday, February 13, 2004 5:19 AM > > To: gda...@li... > > Subject: RE: [Algorithms] (Multi Layer Terrain) Triangle strips for > > "broken" square patches > > > > > Actually, nvtristrip produces great results, but there is > one more > > > problem with it (apart from the slowness), at that is > memory leaks. > > > Seems it doesn't free some of its buffers sometimes. We > > actually had > > > to disable it in order to enable batch preprocessing at all. > > > Otherwise, we cannot process a large list of meshes at > > once, because > > > we run out of memory. :/ I would be happy to find out that > > the problem > > > is on our side, but after some time checking, we couldn't > > find anything. Are you sure it works ok for you? > > > > I was JUST finishing a long whiney message along the lines of "is > > nvtristrip leaking like a sieve, or am I an idiot?". Whew, glad to > > find out that maybe after all it's not me :-) > > > > NVTriStrip is slow, but it leaks so many small blocks, that > it takes > > more time for our debug memory manager to print out the leak alerts > > than for the actual stripification... > > > > > > > > > > ------------------------------------------------------- > > SF.Net is sponsored by: Speed Start Your Linux Apps Now. Build and > > deploy apps & Web services for Linux with a free DVD > software kit from > > IBM. Click Now! > http://ads.osdn.com/?ad_id=1356&alloc_id=3438> &op=click > > > _______________________________________________ > > > GDAlgorithms-list mailing list > > GDA...@li... > > https://lists.sourceforge.net/lists/listinfo/gdalgorithms-list > > Archives: > > http://sourceforge.net/mailarchive/forum.php?forum_id=6188 > > > > > > > ------------------------------------------------------- > SF.Net is sponsored by: Speed Start Your Linux Apps Now. > Build and deploy apps & Web services for Linux with > a free DVD software kit from IBM. Click Now! > http://ads.osdn.com/?ad_id56&alloc_id438&op=click > > _______________________________________________ > GDAlgorithms-list mailing list GDA...@li... > https://lists.sourceforge.net/lists/listinfo/gdalgorithms-list > Archives: http://sourceforge.net/mailarchive/forum.php?forum_ida88 > > > > ------------------------------------------------------- > SF.Net is sponsored by: Speed Start Your Linux Apps Now. > Build and deploy apps & Web services for Linux with > a free DVD software kit from IBM. Click Now! > http://ads.osdn.com/?ad_id=1356&alloc_id=3438> &op=click > > _______________________________________________ > > GDAlgorithms-list mailing list GDA...@li... > https://lists.sourceforge.net/lists/listinfo/gdalgorithms-list > Archives: http://sourceforge.net/mailarchive/forum.php?forum_id=6188 > |